INA121PA
Product Specifications
INA121PA
Texas Instruments — Power Management & Signal Chain
Product Overview
The INA121PA from Texas Instruments is a power management / signal chain IC delivering high efficiency, precision regulation, and robust protection. Texas Instruments (TI) is a global leader in analog ICs, power management solutions, and embedded processors.
Superb Automation partners with leading brands including Qorvo, TI, ADI, XILINX, NVIDIA, Intel, ST, MACOM, Renesas, and Coilcraft to provide complete BOM and supply-chain solutions.
Key Features
- High-efficiency power conversion with low quiescent current
- Wide input voltage range with precision output regulation
- Integrated over-current, over-temperature, and short-circuit protection
- Small footprint packaging for space-constrained designs
- AEC-Q100 qualified options available for automotive applications
- Flexible sequencing and soft-start control
Key Specifications
| Input Voltage | V |
| Output Voltage | V |
| Output Current | A |
| Switching Frequency | kHz |
| Efficiency | % |
| Quiescent Current | µA |
| Topology | — |
| Package | — |
| Operating Temp | °C |
| Protection | — |
Applications
Request Quote — In Stock, 100% New & Original, Full Warranty
INA121PA: the A grade of the FET input amplifier in an eight pin DIP
The INA121PA is the A grade of the plastic DIP member of this family, and the grade letter is the whole difference from the plain P part. Both use the same die, the same eight pin pinout and the same 50 piece tube; the A suffix marks the tighter of the two laser trimmed offset bins. On a channel with a 1 mV full scale the grade decides whether the offset is a small correction or the dominant error, and it costs nothing on the board because the part is pin compatible.
The rest of the family specification is unchanged: 4 pA bias current, 450 µA quiescent, gain from 1 to 10,000 with one resistor, 106 dB of common mode rejection, and inputs that survive 40 V. The part is chosen when a high impedance source has to be measured by a circuit that also has to run from a small battery.
The INA121PA rows as the addendum publishes them
| Part class | FET input instrumentation amplifier, current feedback |
| Vendor status | ACTIVE, production data |
| Data sheet | SBOS078, shared with the plain P row |
| Grade | A suffix, the tighter laser trimmed offset bin |
| Gain equation | G = 1 + 50 kΩ divided by RG |
| Input offset voltage | Plus or minus 200 µV, family row |
| Offset drift | Plus or minus 2 µV per degree |
| Input bias current | Plus or minus 4 pA |
| Common mode rejection | 106 dB |
| Input noise at 1 kHz | 20 nV per root Hz at a gain of 100 |
| Quiescent current | 450 µA |
| Supply range | Plus or minus 2.25 V to plus or minus 18 V |
| Body | PDIP, P outline, 8 pins |
| Carrier | 50 pieces per tube, rails |
| Same die as | INA121P, INA121U and INA121UA |
| Moisture level | Not applicable to this package type |
| Temperature range | Minus 40 to plus 85 degrees |
| RoHS | Yes on the addendum row |
| Lead finish | Published as call TI rather than a fixed finish |
The rows worth a second look are the last three. A tin finish and a nickel palladium gold finish are both possible under a call TI line, and the difference matters for a socketed part that is inserted and removed during calibration: a gold finish survives the abrasion, a tin finish does not after a few dozen cycles. A programme that intends to socket the part should ask for the finish in writing rather than assume it.
What the grade buys on a small signal channel
An offset bin is only meaningful against the full scale of the channel. On a 10 V output the 200 µV row is 20 parts per million and never discussed. On a 1 mV full scale from a thermocouple or a strain bridge it is 20 per cent of the range, and the channel either trims it or lives with it. Trimming at the input side scales with gain; trimming at the output side does not. The difference is worth understanding before the trim circuit is drawn, because a trim that is applied at the wrong node will need to be redone every time the gain resistor changes.
Drift is the part of the error that the trim cannot fix. At 2 µV per degree a 40 degree excursion adds 80 µV, which on a 1 mV full scale is 8 per cent. A channel that has to hold 1 per cent over temperature on that scale is not achievable with this part at any grade, and the honest answer is that the part is the wrong choice rather than that the grade is wrong. On a 100 mV full scale the same drift is 0.08 per cent and the part is comfortable.
Bias current, cable leakage and the guard that is not a pin
At 4 pA of bias current the limiting error is usually the board rather than the amplifier. A fingerprint across two input traces is a few hundred megohms, and at 300 MΩ a 1 V difference between the traces drives 3 nA, which is a thousand times the bias current. Cleaning, a guard ring driven from the reference pin and a conformal coat over the input network are the standard measures, and they cost more engineering time than the part costs money. A board that measures a high impedance source without them will show a drift that moves with humidity and looks like a failing amplifier.
Cable leakage is the second path. A PTFE insulated cable is the usual choice for a picoamp input, and a PVC cable is not, because PVC absorbs moisture and its insulation resistance falls by orders of magnitude in a humid room. The cable shield should be driven from the reference, not grounded, so that the shield to input capacitance is not charged by the signal. These are not optional refinements on a 4 pA part; they are the difference between a channel that works and one that reads noise.
Where this part fits and where it does not
The classic application is a pH or ion selective electrode, where the source impedance is hundreds of megohms and the signal is tens of millivolts. The 4 pA bias current is the enabling number there, and the 20 nV per root Hz noise at a gain of 100 is irrelevant because the source noise dominates. The second application is a piezoelectric or pyroelectric sensor, where the source is capacitive and the bias current sets the droop rate on the sensor capacitance.
Where the part does not fit is a low impedance bridge with a long cable in an industrial cabinet. At a 350 Ω bridge the source resistance is low, the bias current error is negligible, and a bipolar input amplifier with a microvolt offset and a lower noise floor is the better choice on every axis. The FET input stage exists to solve a source impedance problem, and it should be chosen for that problem rather than for the word FET on the data sheet.
Buying INA121PA
The part is active in a 50 piece tube. The purchase order should carry the A suffix because the plain part is the same price band and the wrong grade is not visible on the body without reading the marking. A receiving inspection that photographs the tube label and reads the marking is the cheapest control against a mixed shipment, and it takes a minute per tube.
The second consideration is the second source. A DIP FET input amplifier is a shrinking category, and a programme that runs for a decade should have a written plan for the day the tube cannot be found. Our last time buy and lifetime buy page covers how the buffer quantity is calculated, and the date code management page covers how the stock is rotated so that a buffer does not become the failure.
Questions that arrive with this part number
Whether the A grade is worth the premium over the plain part depends entirely on the full scale of the channel, and that is a question for the customer's specification rather than for a catalogue. Whether the part can be used at a gain of 1 is the second question, and it can: the gain equation gives 1 when RG is open, and the amplifier is stable at unity gain. What changes at unity gain is the common mode range, which is widest at low gain and narrows as the gain rises.
The third question is the supply. Plus or minus 2.25 V is the operational floor, and the part is not specified below it, so a 3.3 V single supply circuit has to generate a negative rail or use a different family. A charge pump that makes minus 3 V from a 5 V rail is the usual answer, and it costs more board area than the amplifier itself.